1. Field of the Invention
The present invention relates to a method and a system for filtering noises, and more particularly to a method and a system for filtering noises in an image scanned by charged particles.
2. Description of the Prior Art
How to inspect defects on a semiconductor device is one of the major subjects in a semiconductor manufacturing process. A prior art method for inspecting defects includes capturing a scanned image of a semiconductor device, such as a scanning electron microscope (SEM) image, and then determining the presence of a defect of the semiconductor device in the scanned image. Therefore, it is very important to obtain an acceptable scanned image.
Referring to FIG. 1, a prior art method for scan imaging comprises scanning primary charged particles 101 across a surface of a sample, and detecting secondary charged particles 102 emitted from the surface of the sample being bombarded by the primary charged particles 101 to form a scanned image accordingly. However, backscattered charged particles 103 are also generated and detected, resulting in noises in scanning process. Traditionally, only one algorithm or noise model is applied to the scanned image to filter noise. However, for different materials on a semiconductor wafer, different noises are generated. Moreover, referring to FIG. 2, the probe spot 20 of the charged particles scans different patterns 22a and 22b along scanning paths 21a, 21b, and 21c or with different scanning directions, which results in different appearances of generated noises. Therefore, the prior art method using only one algorithm or noise model to filter noise cannot obtain a scanned image with better quality. Further, the traditional method for filtering noises is mostly for general electronic images, but is not applicable to an SEM image of wafer with specific characteristics.
Accordingly, it is highly desirable to filter different type noises to obtain a better quality image.